Routine householder applications can represent up to 70% of planning officers’ workloads. Although these cases are often straightforward, they can consume a disproportionate amount of staff time.
To streamline this work, Google DeepMind has developed PlanAssist in collaboration with the UK Government, Google Cloud and Faculty. The tool provides automated support for processing planning applications and preparing assessment recommendations.
PlanAssist first checks whether applicants have submitted the required documents and architectural drawings. It then assesses applications against heritage constraints, including tree preservation orders, as well as local and national planning policies. The system also identifies previously raised issues or objections, such as concerns from local residents.
Using this information, PlanAssist prepares reports with cited recommendations. Planning officers can review and edit the reports before making a final decision.
The tool has been piloted since June by the councils of Barnet, Camden and Dorset. Early results indicate that it has more than halved the time officers spend assessing each application. More than 70% of the reports produced, including checks on compliance with local planning rules and document requirements, were approved by officers without edits.
Sources: Google DeepMind Blog.